Can LLMs Write Correct and Efficient GPU Communication Code?
☆67Jul 7, 2026Updated 4 months ago
Alternatives and similar repositories for CommBench
Users that are interested in CommBench are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- ☆11Oct 29, 2026Updated 2 weeks ago
- mKernel: fast multi-node, multi-GPU fused kernels☆309Oct 17, 2026Updated 3 weeks ago
- Code repository for the SOSP'25 paper DCP: Addressing Input Dynamism In Long-Context Training via Dynamic Context Parallelism.☆23Nov 28, 2025Updated 11 months ago
- ☆54Oct 22, 2026Updated 3 weeks ago
- ☆18Sep 27, 2026Updated last month
- Virtual machines for every use case on DigitalOcean • AdGet dependable uptime with 99.99% SLA, simple security tools, and predictable monthly pricing with DigitalOcean's virtual machines, called Droplets.
- [MLSys 26] 🥇 Solution for Gated Delta Net Track of MLSys 26 Flash infer competition☆38May 22, 2026Updated 5 months ago
- SC'25 UltraAttn: Efficiently Parallelizing Attention through Hierarchical Context-Tiling☆16Aug 14, 2025Updated last year
- htop-like TUI for real-time RDMA network monitoring.☆129Updated this week
- KernelBenchX: A Comprehensive Benchmark for Evaluating LLM-Generated GPU Kernels☆54Jun 1, 2026Updated 5 months ago
- ☆27Feb 27, 2026Updated 8 months ago
- Tinker ↔ KernelBench Integration enabling RL for GPU Kernel Generation☆30Mar 5, 2026Updated 8 months ago
- DINT: Fast In-Kernel Distributed Transactions with eBPF☆52Jul 6, 2024Updated 2 years ago
- Production-ready MoE load balancing via real-time expert replication☆263Oct 15, 2026Updated 3 weeks ago
- ☆27Jan 18, 2026Updated 9 months ago
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- UCCL is an efficient communication library for GPUs, covering collectives, P2P (e.g., KV cache transfer, RL weight transfer), and EP (e.g…☆1,580Updated this week
- DeDe (OSDI '25): an optimization framework for large-scale resource allocation☆15May 18, 2026Updated 5 months ago
- Vortex: Programmable Sparse Attention for Agents as Algorithm Designers☆69Aug 22, 2026Updated 2 months ago
- Mixture-of-experts (MoE) training megakernel for NVL72s☆642Oct 1, 2026Updated last month
- ☆13Sep 3, 2026Updated 2 months ago
- Building the Virtuous Cycle for AI-driven LLM Systems☆307Sep 25, 2026Updated last month
- CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.☆560Updated this week
- An RDMA skew-aware key-value store, which implements the Scale-Out ccNUMA design, to exploit skew in order to increase performance of dat…☆19Jul 1, 2021Updated 5 years ago
- Resource Multiplexing in Tuning and Serving Large Language Models (USENIX ATC 2025)☆10Apr 21, 2026Updated 6 months ago
- Bare Metal GPUs on DigitalOcean Gradient AI • AdPurpose-built for serious AI teams training foundational models, running large-scale inference, and pushing the boundaries of what's possible.
- A benchmark of real-world DL kernel problems☆336Jul 15, 2026Updated 3 months ago
- Venus Collective Communication Library, supported by SII and Infrawaves.☆153Sep 3, 2026Updated 2 months ago
- Some funny cute/cuteDSL code snippets☆35Mar 2, 2026Updated 8 months ago
- ☆47Aug 3, 2026Updated 3 months ago
- A lightweight CuTe-based CUDA kernel for DuoAttention, optimized for large language model inference.☆12Dec 25, 2025Updated 10 months ago
- Automated High-Performance GPU Kernel Generation☆144Jun 1, 2026Updated 5 months ago
- Efficient Long-context Language Model Training by Core Attention Disaggregation☆107Apr 7, 2026Updated 7 months ago
- Mini-SGLang is a mini version of SGLang project ().☆13Sep 24, 2025Updated last year
- An agent for CUDA compute-communication kernel co-design☆37May 7, 2026Updated 6 months ago
- Serverless GPU API endpoints on Runpod - Get Bonus Credits • AdSkip the infrastructure headaches. Auto-scaling, pay-as-you-go, no-ops approach lets you focus on innovating your application.
- Communication patterns for AI, built on top of NCCL device and host APIs☆77Updated this week
- High-performance distributed data shuffling (all-to-all) library for MoE training and inference☆127Mar 7, 2026Updated 8 months ago
- Code and reproducibility artifacts for reconstructing megakernel persistence with CUDA PDL☆27Aug 5, 2026Updated 3 months ago
- ☆136Nov 11, 2024Updated 2 years ago
- Reading seminar in Harvard Cloud Networking and Systems Group☆16Aug 29, 2022Updated 4 years ago
- ☆17Jun 15, 2026Updated 4 months ago
- Tile-based language built for AI computation across all scales☆197Aug 25, 2026Updated 2 months ago
Are these results useful?